Defrosting in the refrigeration industry is used to remove the frost layer accumulated on the evaporators after a period of running time. It is one way to improve the energy efficiency of refrigeration systems. There are many studies about the defrosting process but none of them use computational fluid dynamics (CFD) simulation. The purpose of this thesis is (1) to develop a defrost model using the commercial CFD solver FLUENT to simulate numerically the melting of frost coupled with the heat and mass transfer taking place during defrosting, and (2) to investigate the thermal response of the evaporator and the defrost time for different hot gas temperatures and frost densities. A 3D geometry of a finned tube evaporator is developed and meshed using Gambit 2.4.6, while numerical computations were conducted using FLUENT 12.1. The solidification and melting model is used to simulate the melting of frost and the Volume of Fluid (VOF) model is used to render the surface between the frost and melted frost during defrosting. A user-defined-function in C programming language was written to model the frost evaporation and sublimation taking place on the free surface between frost and air. The model was run under different hot gas temperatures and frost densities and the results were analyzed to show the effects of these parameters on defrosting time, input energy and stored energy in the metal mass of the evaporator. The analyses demonstrate that an optimal hot gas temperature can be identified so that the defrosting process takes place at the shortest possible melting time and with the lowest possible input energy.